Colloquium Computing Science



Date:                            Monday, June 9 2008  

Speaker:                      Prof. Dr.-Ing. Til Aach,

                                    Institute of Imaging and Computer Vision
                                    RWTH Aachen University

Room and time:           5161.0267 (Bernoulliborg), 16.00




Title:    Analysing Superimposed Oriented Patterns




Estimation of local orientation in images is posed of finding the
minimum graylevel variance axis in a local neighborhood. In

bivariate images, the solution is given by the eigenvector

corresponding to the smaller eigenvalue of a 2*2 tensor.

Ideally, the tensor is rank-deficient, i.e., the smaller eigenvalue
vanishes. A large minimal eigenvalue signals more than one

orientation,caused by, e.g., superpositions, crossings, or corners.

We discuss a framework for estimating superimposed orientations,

based on the eigensystem analysis of suitably extended tensors.

The eigensystem analysis does then not directly yield the orientations.

Instead, it provides mixed orientation parameters. We discuss how to

decompose the mixed orientation parameters into the individual
orientations, and derive an invariant feature for describing local
neighborhoods. Applications are, e.g., in texture analysis, directional
filtering and interpolation, feature extraction for corners and crossings,
tracking, and signal separation.



Colloquiumcoördinator is Prof.dr. M. Aiello, Dr. M. Biehl (e-mail: